How Do e-Nutrition Literacy and Faith Shape Positive Nutrition Attitudes? A Machine Learning Approach in Türkiye

Abstract

Background/Objectives: Evidence on religiosity, religious affiliation, and e-nutrition literacy in shaping nutrition attitudes is limited in adult majority-Muslim contexts. The aim of this study is to examine the independent and interactive associations of religiosity, religious affiliation, and e-nutrition literacy with positive nutrition attitudes among adults in T & uuml;rkiye. Methods: This study involved a cross-sectional online survey conducted November-December 2024 via convenience and snowball sampling. After quality checks, 1104 adults remained (mean age = 25.7 years, mean BMI = 23.5 kg/m(2); 69.3% female, 90.7% Muslim). Religiosity was measured with the Duke University Religion Index, and nutrition literacy and positive nutrition attitudes with validated scales. Demographics and anthropometrics were self-reported. Positive Nutrition Attitudes was the primary outcome, predicted by e-nutrition literacy, analyzed using robust OLS and explored for nonlinearities/interactions with Random Forests and SHAP. A generalized linear model tested three-way interactions of e-nutrition literacy, religious affiliation, and religiosity, adjusting for age and BMI. Performance used train or test splits and five-fold cross-validation. Results: e-Nutrition literacy was the strongest predictor (beta = 0.155, p < 0.001). Cross-validated R-2 was modest (about 0.04). Random Forests slightly improved fit (test R-2 about 0.064). SHAP indicated a literacy threshold near 26.1 with predominantly positive contributions above this value. In threshold-stratified models, religiosity showed a positive association (beta = 0.332, p = 0.010). Non-Muslims had higher unadjusted means, but affiliation effects were not significant after adjustment. Conclusions: The results highlight the threshold-dependent role of e-nutrition literacy in positive nutrition attitudes and the independent effect of religiosity. These results suggest that boosting literacy above the critical threshold and incorporating religious values may support healthier nutrition behaviors.